ESICT
The project is closed.
The ESICT-consortium (Experience-oriented Sharing of health knowledge via Information and Communication Technology) received a grant of 10.7 mill. D.kr. from the Strategic Research Council. The grant is for research in methods and development of technologies providing citizens with an innovative information system on health and disease based on information technology, language technology and formalized medical knowledge. The ESICT system allows users to ask questions in natural language which will be answered the same way.
In our society we need healthcare IT systems that integrate patients, relatives and health professionals as equitable partners. Such systems should grant users access to information whenever they are confronted with diagnoses, procedures, etc. or when they feel a need of discussing a health or disease related issue. Free and equal access to knowledge regardless of social background, age or knowledge level is being perceived more and more as an essential prerequisite to the vision of health for all. In response to this, the ESICT project has an experience-oriented approach, as it focuses on the creation of a customized dialogue adapted to the needs of a diverse user population. The language used is Danish, but the methods are generic and adaptable to other languages.
The main result of ESICT will be a prototype of a QA system covering issues related to diabetes and ischemic heart disease. It will give individualized feedback, depending on the context in which the particular question is asked. This prototype can be deployed right away by industry or information system providers who want to focus on one of these diseases.
When we have demonstrated the potential of the system within the selected scope, the prototype can be further developed for other disease or health topics. Further, our approach naturally extends to other languages and even to other domains where a similar termbase and ontology exist. It will enable the involved project partners to develop systems which can be exported to other countries.
The prototype will incorporate a truly high precision, relatively robust parser, high precision database querying and generation of useful, unambiguous answers and will introduce new techniques in natural language processing. The combination of ontology navigation and extraction of information from other sources is highly innovative. This research will also be useful in other applications where syntactic analysis is used in complicated pipelines to improve output quality, including automatic summarization, information search and multilingual applications.
Technical University of Denmark
Søren Brunak, Center for Biologisk Sekvensanalyse
Sorano limited
Ulrich Andersen, Center for Biological Sequence Analysis
Medical University Center Freiburg
Csaba Huszka, Institute of Medical Biometry and Medical Informatics
Medical University of Graz
Stefan Schulz, Institute for Medical Informatics, Statistics und Documentation
Contact
Bente Maegaard
Centre for Language Technology
University of Copenhagen